Dense wavelength division multiplexed (DWDM) optical networks increase their transmission capacity by employing multiple co-propagating, discrete, wavelength channels, each carrying independent data streams. Broadband fiber optic devices, such as variable attenuators, couplers, and switches having a controllable spectral response, are critical components of DWDM systems. Currently, DWDM systems operate in the 1550 nm spectral region because of the availability of optical amplifiers containing erbium-doped optical fibers. However, as amplifier technology develops, and capacity demands increase, DWDM systems are expected to expand their spectral extent and increase their channel density.
Optical power, as it propagates in a single-mode optical fiber, or any other waveguide or bulk material, experiences dispersion, i.e. differing wavelengths propagate at different speeds. In an optical fiber, modal extent and phase velocity are affected by both the dispersion of the material (material dispersion) and the dispersion of the waveguide (waveguide dispersion) causing the light to pass through at different speeds. Thus, across a given wavelength region, differences between the dispersions of the material and waveguide through which light propagates can result in nonuniform spectral performance of fiber-based devices.
Dispersion is often represented in terms of a material's refractive index (n) as a function of optical wavelength (.lambda.), i.e. as n(.lambda.). In dispersive materials, the refractive index of the material changes with wavelength. The relevant parameter when describing modal dispersion or multimode distortion in optical fibers is the effective mode index, also referred to herein as "effective mode dispersion", n.sub.eff (.lambda.), which, in simple waveguide geometries, can be calculated using the material dispersion of the fiber's cladding and core, n.sub.clad (.lambda.) and n.sub.core (.lambda.), respectively, and geometric parameters. Sellmeier dispersion equations for the cladding and core in a single mode optical fiber are provided by J. Gowar, in Optical Communication Systems, ch. 3, 58-77 (2d ed.993). For a glass fiber, the material dispersions for the cladding and core are calculated from the following Sellmeier equations (1a) and (1b), respectively, which are valid from 0.3-3.0 .mu.m: ##EQU1##
The dispersions of the cladding, n.sub.clad (.lambda.) and core, n.sub.core (.lambda.), and the effective mode dispersion, n.sub.eff (.lambda.) for a silica glass optical fiber having a core with a slightly raised refractive index relative to the surrounding cladding are plotted in FIG. 1. Although all materials are dispersive to some extent, a hypothetical material exhibiting no dispersion would be represented in the graph of FIG. 1 as a horizontal line. The greater the dispersion, the steeper the slope (negative or positive). As used herein, the term "dispersion" refers to the slope of the line formed from a plot of a material's change in refractive index versus change in wavelength. As can be seen from the slope of n.sub.eff (.lambda.) in FIG. 1, a single mode optical fiber is dispersive.
Because the effective mode index is dispersive, fiber-based devices may exhibit spectrally non-uniform performance, which is undesirable for many broadband device applications. An example of this is a side-polished fiber (SPF)-based attenuator. Cargille Refractive Index Liquids, which may be coupled onto the attenuator, each have a well-characterized refractive index, n.sub.D, where subscript D denotes the Sodium D-Line wavelength (.lambda.=589 nm), and a well-characterized dispersion curve. As disclosed in copending commonly assigned U.S. application Ser. No. 09/026,755 entitled "Fiber Optic Attenuators and Attenuation Systems", the disclosure of which is incorporated herein by reference, placing a coupling oil (n.sub.D =1.456 at 27.9.degree. C.) on a SPF coupler induces power loss (attenuation). FIG. 2a is a plot of attenuation (optical energy transmission) in decibels versus wavelength (1520-1580 nm) for a SPF coupler having a 95% polished cladding level. As shown in FIG. 2a, the attenuation is not uniform across the spectral region. This spectral nonuniformity is observed because the dispersion of the oil, n.sub.oil (.lambda.), which is calculated from equation (2) as ##EQU2##
(where T is the temperature) is mismatched to that of the fiber, n.sub.eff (.lambda.). This dispersion mismatch is depicted in FIG. 2b, where the slope of n.sub.oil (.lambda.) differs from that of n.sub.eff (.lambda.). By contrast, if the dispersion of the oil matched that of the fiber, the graphic representations of the corresponding dispersions would be approximately parallel, and the attenuation would be almost constant or substantially uniform across the wavelength band with only small variations being observed.
As disclosed in the aforementioned U.S. application Ser. No. 09/026,755, certain organic polymers having an index of refraction close to that of the fiber can be applied to the exposed surface of a SPF optic for use in variable optical attenuators (as described below) Such polymers exhibit a change in refractive index proportional to a change in temperature. OPTI-CLAD.RTM. 145, which is available from Optical Polymer Research, Inc. is an example of such a polymer. Although the refractive index of such organic polymer materials can be altered at a given wavelength to match that of the fiber, the use of known polymers is limited in broadband applications because of the dispersion mismatch between the polymer and the fiber across the wavelength band of interest.
Control over the spectral response of a broadband fiber optic device can be strongly dependent on the dispersion mismatch between the fiber and any coupling materials present. Therefore, polymer formulations are desirable that would allow not only control of the refractive index of the polymer overlying the optical fiber, but also control of the dispersion properties of the polymer. Such dispersion control would permit correction of the polymer's dispersion to substantially match that of the fiber and, alternatively, would allow the dispersion of the polymer to be controllably altered from that of the fiber. Such dispersion controllable materials would be useful in broadband applications, such as in the 1500-1600 nm region, where control of spectral response is important. In addition, dispersion controllable polymer materials would be useful in the fabrication of many broadband fiber optic devices, such as variable optical attenuators (VOAs), couplers, and switches.